Introduction to Bioinformatics - SJSUkhuri/Yverdon_2010/Yverdon_2010_MSA.pdf · Introduction to...

July 2010

4.1©2010 Sami Khuri

Yverdon Les Bains

Introduction to Bioinformatics

@2010 Sami Khuri

Introduction to

BioinformaticsSami Khuri

Department of Computer Science

San José State University

San José, California, USA

[email protected]

www.cs.sjsu.edu/faculty/khuri

©2010 Sami Khuri

Multiple Sequence Alignment

� Progressive Alignment

� Iterative Pairwise

� Guide Tree

� ClustalW

� Co-linearity

� Multiple Sequence

Alignment Editors

July 2010


Yverdon Les Bains


©2010 Sami Khuri

* 20 * 40 * 60 * 80

Wombat : AAAGTTAATGAGTGGTTATCCAGAAGTAGTGACATTTTAGCCTCTGATAACTCCAACGGTAGGAGCCATGAGCAGAGCGCAGA : 83

Opossum : AAAGTTAATGAGTGGTTATTCAGAAGTAATGACGTTTTAGCCCCAGATTACTCAAGTGTTAGGAGCCATGAACAGAATGCAGA : 83

Armadillo : AAAGTTAACGAGTGGTTTTCCAGAGGTGATGACATATTAACTTCTGATGACTCACACGATAGGGGGTCTGAATTAAATGCAGA : 83

Sloth : AAAGTTAATGAGTGGTTTTCCAGAAGTGATGACATACTAACTTCTGATGACTCACACAATGGGGGGTCTGAATCAAATGCAGA : 83

Dugong : AAAGTTAATGAGTGGTTTTTCAGAAGTGATGGCCTG---------GATGACTTGCATGATAAGGGGTCTGAGTCAAATGCAGA : 74

Hyrax : AAAGTTAATGAGTGGTTTTCCAGAAGTGACAACCTA---------AGTGATTCACCTAGTGAGGGGTCTGAATTAAATGGAAA : 74

Aardvark : AAAGTTAATGAGTGGTTTTCCAGAAGTGATGGCCTG---------GATGGCTCACATGATGAAGGGTCTGAATCAAATGCAGA : 74

Tenrec : AAGGTTAACGAGTGGTTTTCCAAAAGCCACGGCCTG---------GGTGACTCTCGCGATGGGCGGCCTGAGTCAGGCGCAGA : 74

Rhinoceros : AAAGTTAATGAGTGGTTTTCCAGAAGTGATGAAATATTAACTTCTGATGACTCACATGATGGGGGGCCTGAATCAAATACTGA : 83

Pig : AAAGTTAATGAGTGGTTTTCTAGAAGCGATGAAATGTTAACTTCTGACGACTCACAGGACAGGAGGTCTGAATCAAATACTGG : 83

Hedgehog : AAAGTGAATGAATGGCTTTCCAGAAGTGATGAACTGTTAACTTCTGATGACTCATATGATAAGGGATCTAAATCAAAAACTGA : 83

Human : AAAGTTAATGAGTGGTTTTCCAGAAGTGATGAACTGTTAGGTTCTGATGACTCACATGATGGGGAGTCTGAATCAAATGCCAA : 83

Rat : AAAGTGAATGAGTGGTTTTCCAGAACTGGTGAAATGTTAACTTCTGACAATGCATCTGACAGGAGGCCTGCGTCAAATGCAGA : 83

Hare : AAAGTTAACGAGTGGTTCTCCAGAAGTAATGAAATGTTAACTCCTGATGACTCACTTGACCGGCGGTCTGAATCAAATGCCAA : 83

AAaGTtAAtGAgTGGtTtTccAgAagt atga T gatgactca gat g gg cTga t aaatgc ga

* 100 * 120 * 140 *

Wombat : GGTGCCTAGTGCCTTAGAAGATGGGCATCCAGATACCGCAGAGGGAAATTCTAGCGTTTCTGAGAAGACTGAC : 156

Opossum : GGCAACCAATGCTTTAGAATATGGGCATGTAGAGACA---GATGGAAATTCTAGCATTTCTGAAAAGACTGAT : 153

Armadillo : AGTAGCTGGTGCATTGAAAGTT------TCAAAAGAAGTAGATGAATATTCTAGTTTTTCAGAGAAGATAGAC : 150

Sloth : AGTAGTTGGTGCATTGAAAGTT------CCAAATGAAGTAGATGGATATTCTGGTTCTTCAGAGAAGATAGAC : 150

Dugong : AGTAGCTGGTGCTTTAGAAGTT------CCAGAAGAAGTACATGGATATTCTAGTTCTTCAGAGAAAATAGAC : 141

Hyrax : AGTGGCTGGTCCAGTAAAACTT------CCAGGTGAAGTACATAGATATTCTAGTTTTCCAGAGAACATAGAT : 141

Aardvark : AATAGGTGGTGCATTAGAAGTT------TCAAATGAAGTACATAGTTACTCTGGTTCTTCAGAGAAAATAGAC : 141

Tenrec : CGTAGCTGTAGCCTTCGAAGTT------CCAGACGAAGCATGTGAATCTTATAGTTCTCCAGAGAAAACAGAC : 141

Rhinoceros : AGTAGCTGGTGCAGTAGAAGTT------CAAAATGAAGTAGATGGATATTCTGGTTCTTCAGAGAAAATAGGC : 150

Pig : GGTAGCTGGTGCAGCAGAGGTT------CCAAATGAAGCAGATGGACATTTGGGTTCTTCAGAGAAAATAGAC : 150

Hedgehog : AGTAACTGTAACAACAGAAGTT------CCAAATGCAATAGATAGRTTTTTTGGTTCTTCAGAGAAAATAAAC : 150

Human : AGTAGCTGATGTATTGGACGTT------CTAAATGAGGTAGATGAATATTCTGGTTCTTCAGAGAAAATAGAC : 150

Rat : AGCTGCTGTTGTGTTAGAAGTT------TCAAATGAAGTGGATGGATGTTTCAGTTCTTCAAAGAAAATAGAC : 150

Hare : AGTGGCTGGTGCATTAGAAGTC------CCAAAGGAGGTAGATGGATATTCTGGTTCTACAGAGAAAATAGAC : 150

gt gctg tgc t gAagtt cA a gaag a atggatatT t Gtt TtCagAgAA Atagac

Part of the alignment of the DNA sequences of the BRCA1 gene

From “Bioinformatics and Molecular Evolution” by Paul Higgs and Teresa Attwood

©2010 Sami Khuri

* * * * *

Wombat : KVNEWLSRSSDILASDNSNGRSHEQSAEVPSALEDGHPDTAEGNSSVSEKTD : 52

Opossum : KVNEWLFRSNDVLAPDYSSVRSHEQNAEATNALEYGHVET-DGNSSISEKTD : 51

Armadillo : KVNEWFSRGDDILTSDDSHDRGSELNAEVAGALKV--SKEVDEYSSFSEKID : 50

Sloth : KVNEWFSRSDDILTSDDSHNGGSESNAEVVGALKV--PNEVDGYSGSSEKID : 50

Dugong : KVNEWFFRSDGL---DDLHDKGSESNAEVAGALEV--PEEVHGYSSSSEKID : 47

Hyrax : KVNEWFSRSDNL---SDSPSEGSELNGKVAGPVKL--PGEVHRYSSFPENID : 47

Aardvark : KVNEWFSRSDGL---DGSHDEGSESNAEIGGALEV--SNEVHSYSGSSEKID : 47

Tenrec : KVNEWFSKSHGL---GDSRDGRPESGADVAVAFEV--PDEACESYSSPEKTD : 47

Rhinoceros : KVNEWFSRSDEILTSDDSHDGGPESNTEVAGAVEV--QNEVDGYSGSSEKIG : 50

Pig : KVNEWFSRSDEMLTSDDSQDRRSESNTGVAGAAEV--PNEADGHLGSSEKID : 50

Hedgehog : KVNEWLSRSDELLTSDDSYDKGSKSKTEVTVTTEV--PNAIDXFFGSSEKIN : 50

Human : KVNEWFSRSDELLGSDDSHDGESESNAKVADVLDV--LNEVDEYSGSSEKID : 50

Rat : KVNEWFSRTGEMLTSDNASDRRPASNAEAAVVLEV--SNEVDGCFSSSKKID : 50

Hare : KVNEWFSRSNEMLTPDDSLDRRSESNAKVAGALEV--PKEVDGYSGSTEKID : 50

KVNEWfs4 6 d s e n e eki

Alignment of BRCA1 protein sequences for the same region on the gene

Aligning BRCA1 Sequences

From “Bioinformatics and Molecular Evolution” by Paul Higgs and Teresa Attwood

July 2010


Yverdon Les Bains


©2010 Sami Khuri

What is Multiple Alignment

Most simple extension of pairwise alignment

Given:

• Set of sequences

• Match matrix

• Gap penalties

Find:

Alignment of sequences such that an optimal

score is achieved.

©2010 Sami Khuri

Uses of Multiple Alignment

A good alignment is critical for further analysis

• Determine the relationships between a group

of sequences

• Determine the conserved regions

• Evolutionary Analysis

– Determine the phylogenetic relationships and evolution

• Structural Analysis

– Determine the overall structure of the proteins

July 2010


Yverdon Les Bains


©2010 Sami Khuri

Alignment Difficulties

We cannot use sequence comparison algorithms

(dynamic programming) and just add more sequences.

With each sequence, another dimension is added that we

need to search for the optimal alignment.

Add 1

sequence

©2010 Sami Khuri

Hyperlattice Computation

NA

NS

NV

s

NA

-N

NV

s

V

S

A

NA

NS

-N

s

-N

NS

-N

s

-N

-N

NV

s

N

N

N

s

+

A

S

-

δ

+

+

+

+

+

+

+

=

A

-

-

-N

NS

NV

-

S

-

NA

-N

NV

-

-

V

NA

NS

-N

-

S

V

NA

-N

-N

A

-

V

-N

NS

-N

A

S

-

-N

-N

NV

A

S

V

N

N

N

max

NA

NS

NV

δ

δ

δ

δ

δ

δ

δ

s

s

s

s

s

s

s

s

k=3 2k –1=7

July 2010


Yverdon Les Bains


©2010 Sami Khuri

MSA: Exact vs. Heuristic

• The exact algorithm

– traverses the entire search space

– finds overall measure of alignment quality and tries to maximize this quality.

• The operation is computationally intensive.

• The largest computers can only optimally align a few sequences (7-8).

• Therefore, we have to use heuristics; i.e., faster algorithms, if we want to align many sequences.

©2010 Sami Khuri

Heuristic Algorithms

• Based on a progressive pairwise alignment approach– ClustalW (Cluster Alignment)

– PileUp (GCG)

– MACAW

• Builds a global alignment based on local alignments

• Builds local multiple alignments

• Based on Hidden Markov Models

• Based on Genetic algorithms.

July 2010


Yverdon Les Bains


©2010 Sami Khuri

Progressive Strategies for MSA

• A common strategy to the MSA problem is

to progressively align pairs of sequences.

– A starting pair of sequences is selected and

aligned

– Each subsequent sequence is aligned to the

previous alignment.

• Progressive alignment is a greedy

algorithm.

©2010 Sami Khuri

Iterative Pairwise Alignment

• The greedy algorithm:

align some pair

while not done

pick an unaligned string “near”

some aligned one(s)

align with the previously aligned group

• There are many variants to the algorithm.

July 2010


Yverdon Les Bains


©2010 Sami Khuri

ClustalW: Package for MSA• ClustalW [the W is from Weighted] is a software package

for the MSA problem.

• Different weights are given to sequences and parameters in

different parts of the alignment to and create an alignment

that makes sense biologically.

• Scalable Gap Penalties for protein profile alignments

– A gap opening next to a conserved hydrophobic residue can be

penalized more heavily than a gap opening next to a hydrophilic

residue.

– A gap opening very close to another gap can be penalized more

heavily than an isolated gap.

©2010 Sami Khuri

Steps of ClustalW

S1

S2

S3

S4

4

3

2

1

4321

S

4S

74S

494S

SSSS

All Pairwise

Alignments

S1

S3

S2

S4

Distance

Cluster Analysis

Similarity Matrix Dendrogram

Multiple Alignment Step:1. Aligning S1 and S3

2. Aligning S2 and S4

3. Aligning (S1,S3) with (S2,S4).

July 2010


Yverdon Les Bains


©2010 Sami Khuri

ClustalW: An Example

By using the same five

sequences and aligning

them with CLUSTALW,

we get the illustrated

results.

* = identity

: = strongly conserved

. = weakly conserved

©2010 Sami Khuri

July 2010


Yverdon Les Bains


©2010 Sami Khuri

Red Blood Cells

©2010 Sami Khuri

Hemoglobin Structure

• 4 subunits: 2 alpha

and 2 beta (red and

gold)

• Each subunit holds a

heme group

• Each heme group

contains an iron

atom, which is

responsible for the

binding of oxygen

July 2010


Yverdon Les Bains


©2010 Sami Khuri

Practical Considerations

• When to use Clustal?

• Can be used to align any group of protein or nucleic acid sequences that are related to each other over their entire lengths.

• Clustal is optimized to align sets of sequences that are entirely co-linear, i.e. sequences that have the same protein domains, in the same order.

©2010 Sami Khuri

When Not To Use Clustal

• Sequences do not share common ancestry.

• Sequences are partially related.

• Sequences include short non overlapping fragments.

July 2010


Yverdon Les Bains


©2010 Sami Khuri

Structural Alignment

• What you really want to do is “align regions of similar function”.

• These are the areas that are evolutionarily conserved. (Folds, domains, disulfide bonds)

• Problem

– The computer does not know anything about the structure or function of the proteins.

• Solution

– Use computer alignment as a first step, then manually adjust the alignment to account for regions of structural similarity.

©2010 Sami Khuri

MSA Editors

• Once the multiple alignment is produced, it may be necessary to edit the sequence manually to obtain a more reasonable or expected alignment.

• Some of the considerations for an editor:

– the use of colors to aid in the visual representation of the alignment,

– the capability of recognizing the alignment format,

– the ability of using the mouse to add, delete, or move sequences, thus allowing for an adequate windows interface.

July 2010


Yverdon Les Bains


©2010 Sami Khuri

Divide and Conquer Alignment

Use the divide and conquer

technique to perform multiple

sequence alignment.

MSA with the Divide &

Conquer Method by J. Stoye

Gene 211(2), GC45-GC56, 1998.

(Gene-COMBIS).

http://bibiserv.techfak.uni-

bielefeld.de/dca/algorithm/

©2010 Sami Khuri

Appropriate Approaches

Introduction to Bioinformatics - SJSUkhuri/Yverdon_2010/Yverdon_2010_MSA.pdf · Introduction to...

Documents

Transcript of Introduction to Bioinformatics - SJSUkhuri/Yverdon_2010/Yverdon_2010_MSA.pdf · Introduction to...